An evolutionary system for near-regular texture synthesis

Fei Wu,Changshui Zhang,Jingrui He
DOI: https://doi.org/10.1016/j.patcog.2006.10.017
IF: 8
2007-01-01
Pattern Recognition
Abstract:Near-regular texture is probably among the most difficult to handle in the texture synthesis area, because the synthesis must preserve the holistic structural property and the local randomness simultaneously. In this paper, motivated by the relationship between a near-regular texture image and an evolutionary system, we propose a novel texture synthesis algorithm. By defining individuals with appropriate attributes and behaviors, we convert the texture synthesis problem to an evolution process of an evolutionary system. It can achieve high-quality synthesized results on a large variety of near-regular textures without any extra overhead for memory and pretreatment, and the speed approaches real-time. Moreover, it can be easily generalized to deal with other kinds of textures.
What problem does this paper attempt to address?